Artificial intelligence, vanishing legal jobs, and art

by Rob Tiller

Is technology rendering humans obsolete? The answer is, as to some activities, yes. But could it help us better understand our true nature? It could.

Last week the NY Times reported that new computer programs were able to do legal review of electronic documents more accurately and much cheaper than human lawyers. This is a milestone in technology, and one with big implications for the legal profession, and other professions, too.

In my professional capacity at Red Hat as a purchaser of legal services, I’m happy to consider using these money-saving tools. And having as a young lawyer spent hours doing dreary document review, I’m happy to think that humans may be able to hand such drudgery off to computers and do more stimulating things with themselves. But lots of law firms survive and thrive by selling document review services. Automating such work will cause painful dislocations, as many legal jobs go down the tubes.

It’s strange to think of part of lawyering going the way of the gas station attendant. As computer-driven technology replace partially or completely entire categories of work, such as huge swathes of manufacturing, educated professionals have assumed that they were immune. But that is clearly wrong. The triumph of Watson on Jeopardy a couple of weeks ago and the success of legal document review programs shows that more change is on the way.

This is somewhat frightening. But it also forces us to confront the interesting question of what we can usefully do, other than the logic-driven work that computers are now taking over. Since Peraclesian Athens, we’ve assumed that human reasoning was the crowning glory of creation, but we need to revisit that understanding of nature, and human nature.

A few months back I read The Science of Fear, by Daniel Gardner, which offers some interesting thinking on the inherent flaws in human rationality. Gardner focuses on how we systematically underestimate some risks, like the risk of highway accidents, and overestimate others, such as the risk of terrorism and violent crime. Our journalism establishment is heavily invested in promulgating scare stories on such subjects, and we seem in general to like such stories, or at least eat them up. Gardner discusses the psychological basis of this odd characteristic, and the possibility that with more quantitative analysis we could work around the problem. I’m in favor of more careful quantitative analysis of problems, but I doubt that will much affect how human minds work.

David Brooks wrote a surprisingly thoughtful (especially for a conservative) column in the Times this week about human nature. He posited that various kinds of scientists are coming to think of humans are fundamentally social, and that it’s a mistake to think of them as isolated individuals. He also emphasized that our unconscious, emotional capacities are more important than our reasoning. In other words, the way our minds work is mostly non-rational. We aren’t just poorly fashioned reasoning machines, but a different kind of being.

This is worth a lot more exploring. There are various things that humans do that are non-rational, but not unintelligent. Artistic activity is a prime example. When we sing or dance, we’re connecting to our selves and others in a way that is richly human. Telling stories in various media is a constant of our lives. These are things that we as a species are really good at, and we enjoy. They aren’t peripheral to our lives and culture — they’re central. Our computers may get at making art, but they can’t replace us in those activities, because we realize ourselves in them.